From the Ground Up-OPUS Rapid Static
May 1, 2007
I have written several columns over the years about the National Geodetic Survey (NGS) and its importance to the surveying and mapping professions. I don’t hide the fact that I am a big fan of the NGS. This agency is incredibly important to me in my mapping profession, and is equally important to the surveying community.
I conduct research on the NGS website for new project areas and download data sheets for control points on a regular basis. I make use of its Continuously Operating Reference Stations (CORS) network on a daily basis during the spring and fall flying seasons for post-processing airborne GPS data. And I have been using the NGS Online Positioning User Service (OPUS) for several years, both to establish the positions of new control points and as an important quality control step for long ground occupations conducted during aerial photography or LiDAR missions.
I have had the opportunity to check OPUS solutions many times over the last several years and the results have always fit very well with the published positions of the occupied control points. The one major drawback to OPUS has been the requirement for a minimum of two hours of occupation data. That has limited the use of OPUS in some traditional applications due to the loss of productivity in two-hour sessions. But things changed considerably on January 31 of this year when NGS released OPUS Rapid Static, or OPUS-RS. OPUS-RS is designed to provide accurate results with GPS data sets as short as 15 minutes. Dr. Richard Snay, manager of the National CORS Program, emphasized the reasons behind the development of OPUS-RS: “The requirement for two hours of data has limited the use of OPUS. With OPUS-RS, we expect many more professionals to make use of this service. In fact, we anticipate a tenfold increase in the use from the current level of about 15,000 requests per month.” OPUS-RS is proven technology; an operational prototype was used and tested for 15 months prior to its declaration of Initial Operational Capability and release to the surveying and mapping community.
Application of OPUS-RSYou are likely familiar with OPUS. The application of OPUS-RS is very similar to the original OPUS. Access to the service is gained from the NGS web page at www.ngs.noaa.gov. Dialog boxes are provided at the OPUS-RS site for browsing to the appropriate GPS data file on your computer system, specifying the antenna type and antenna height above the mark, and supplying the E-mail address where the results of the OPUS solution will be forwarded. Like the original version, OPUS-RS uses observation data from the CORS network to determine positional data for your observation.
Only dual-frequency data is accepted, and the data must be in either RINEX format or one of several supported popular receiver formats. The minimum amount of data for OPUS-RS is 15 minutes; the maximum is two hours. Longer observations should be submitted to the original OPUS. When all fields are complete in the dialog box, you simply click a button at the bottom of the screen to upload your data. Results are forwarded to the E-mail address that you provided, normally within just a few minutes after submission.
Data submitted to OPUS is processed with respect to three reference stations in the CORS network. OPUS makes the selection of the three sites to be used from a combination of distance from your occupation site, number of observations and site stability, among others. Therefore, the sites selected may not be the nearest to your observation. The algorithm is significantly different for OPUS-RS. NGS explains this algorithm at the OPUS-RS site. OPUS-RS searches the reference stations in order of increasing distance from the user’s station, selecting reference stations that have suitable data. The search stops when either six stations have been selected or the distance exceeds 200 km. Because of the geographic interpolation algorithms used, OPUS-RS will only process a file for which the user’s station is either inside the polygon enclosing the selected reference stations, or no more than 50 km outside that polygon. Also, if the search algorithm does not find at least three stations within 200 km, OPUS-RS will refuse to attempt a solution.
Positional information is provided in International Terrestrial Reference Frame (ITRF) and NAD 83 coordinates as well as UTM and State Plane Coordinate positions. Elevations are provided as both ellipsoidal and orthometric heights. Users can choose the state plane projection for the provided results when submitting data. For example, options for Kentucky include Kentucky north, Kentucky south and the single zone state plane projection.
The results page also provides information on the base stations used in the solution and the distance from the occupation point to each of them. The nearest NGS published control point is also listed.
Accuracy of OPUS-RSRecent experiments and testing by NGS of 15-minute data sets indicate a 92 percent confidence interval of 5 cm for the north and east positions and 10 cm for the ellipsoidal height. Similarly, the 97 percent confidence interval is 10 cm for the north and east positions and 20 cm for the ellipsoidal height. Not bad at all for such short occupations. Accuracies should be expected to increase with longer occupation times.
The results E-mailed at the completion of an OPUS-RS run also provide solution statistics that allow some insight into the expected accuracy. These should be carefully reviewed for all solutions.
The first statistic is the percentage of observations used. With the original OPUS, NGS suggested that a good solution used at least 90 percent of the observations. Secondly, quality indicators are provided. According to NGS, the two quality indicators provided are the average values of the W-RATIO of the last three epochs of data. The W-ratio is a statistical measure for the ambiguity resolution, measuring the separation between the best, and second best candidate sets. Typically the W-ratio increases with time, indicating a better solution. The quality indicator reported by OPUS-RS is the average of this ratio at the last three epochs. The indicators are based on a network mode run and a rover mode run. Higher numbers here are good and quality indicators below 2.0 indicate caution is necessary in the positional results. And finally, the value of the normalized RMS is provided. This is a unitless quantity of the reference standard deviation. A value greater than 1.0 indicates the actual noise in the data was greater than what was indicated by the weights assigned in processing the data. A value less than 1.0 indicates less noise than predicted by the assigned weights.
We at Photo Science were involved in the testing of OPUS-RS last year prior to its public release; we ran several observations over NGS published points. The data always fit within the accuracies defined by NGS. In February, I conducted more extensive tests using the base station that is permanently installed at our office building. The base station software was configured to collect 15- and 30-minute data sets over a weekend. The data files were then submitted to OPUS-RS and the resulting positions compared to the three-dimensional position that we had previously established for the base.
The horizontal position of the base station was established by static ties to three local HARN stations. The elevation of the base was determined by differential level runs to the antenna reference point. All comparisons were made in the Kentucky state plane single zone with orthometric heights based on NAVD 88 in meters. The table above illustrates the results.
Originally, the horizontal position of the base station was referenced to the Kentucky HARN adjustment. Prior to performing these tests, the position of the base was adjusted to the new NAD 83 (NSRS 2007) released by NGS in early February. This adjustment reconciled any differences between the HARN and CORS and provides near perfect consistency between the two. This adjustment resulted in the position of this base station shifting 1 cm in the northing and 5 mm in the easting. This adjustment moved the known position toward the average of the 30 sessions for both the 15- and 30-minute observations performed for the test, thus improving the results in both cases.
Six CORS were used in the OPUS-RS solutions for this test. The distances to these CORS ranged from 68 km to LOU1 located southeast of Louisville to 117 km to KYBO located in Boone County in northern Kentucky. The map on the opposite page illustrates the locations of these CORS in relation to the base station used for the test. The concentric circles in the graphic represent radii of 50, 100 and 150 km from the base station.
The results of the test are consistent with NGS’ accuracy predictions. In fact, the horizontal accuracy is significantly better than the stated accuracies. You will also notice that there was a very slight improvement in accuracy when the sessions were extended from 15 to 30 minutes.
Free Tools for GPS OccupationsSnay described the importance of this service to the surveying profession: “OPUS-RS provides direct access to the National Spatial Reference System (NSRS) in a way much easier than ever before. Considering that time is money, the importance of accurate positioning with short observations is very clear.” He also discussed a new application of OPUS-DB (database) that should be available in a few months. “OPUS-DB will allow surveying professionals to easily share their work with others in the professional community. Registered users will be able to submit a minimum of four hours of observation data, photos and descriptions of the new occupation, and log materials of observation procedures. Positional information for the new point will be available within our data sheet search engine once an acceptable OPUS solution is confirmed.”
Once again, NGS has provided surveying and mapping professionals with a significant new tool to help us when working with GPS occupations. And this tool is free to the professional community. The original OPUS continues to have important applications for GPS occupations of two hours or more, and OPUS-RS provides significant accuracy for short occupations, thereby increasing our productivity and decreasing project costs.